fix missing fields when ordering by

[isinyaaa]

Description

Related: kubeflow/model-registry#358

How Has This Been Tested?

Tested locally by deploying with kind. Attempted to order various context queries.

from model_registry import ModelRegistry
import contextlib

mr = ModelRegistry("http://localhost", 8080, author="me", is_secure=False)

with contextlib.suppress(Exception):
    mr.register_model(
        "model1",
        "s3",
        version="1",
        model_format_name="1",
        model_format_version="2",
    )
with contextlib.suppress(Exception):
    mr.register_model(
        "model",
        "s3",
        version="2",
        model_format_name="1",
        model_format_version="2",
    )
with contextlib.suppress(Exception):
    mr.register_model(
        "model",
        "s3",
        version="3",
        model_format_name="1",
        model_format_version="2",
    )
with contextlib.suppress(Exception):
    mr.register_model(
        "model2",
        "s3",
        version="1",
        model_format_name="1",
        model_format_version="2",
    )
with contextlib.suppress(Exception):
    mr.register_model(
        "model",
        "s3",
        version="1",
        model_format_name="1",
        model_format_version="2",
    )


print("Registered models by id")
for model in mr.get_registered_models().limit(1):
    print(model)

print("Registered models by create")
for model in mr.get_registered_models().order_by_creation_time().descending().limit(1):
    print(model)

print("Registered models by update")
for model in mr.get_registered_models().order_by_update_time().limit(1):
    print(model)


print("model versions by id")
for model in mr.get_model_versions("model").limit(1):
    print(model)

print("model versions by create")
for model in (
    mr.get_model_versions("model").order_by_creation_time().descending().limit(1)
):
    print(model)

print("model versions by update")
for model in mr.get_model_versions("model").order_by_update_time().limit(1):
    print(model)

Merge criteria:

• The commits are squashed in a cohesive manner and have meaningful messages.
• Testing instructions have been added in the PR body (for PRs involving changes that are not immediately obvious).
• The developer has manually tested the changes and verified that the changes work

[tarilabs]

/lgtm

thanks @isinyaaa !

Would be awesome to consider how to run the tests always to ensure non-regressions (I'm not saying that has to be done as part of this PR)

[isinyaaa]

@tarilabs updated the tag from "Fixes" to "Related", so the issue won't close automatically, wdyt? can we merge this?

[tarilabs]

thanks @isinyaaa yes I believe keeping the link this way (related) is more appropriate.
once merged, kindly double-check if the steps in the productization process are followed automatically by devops, or further actions are needed, thanks!

[tarilabs]

while we're not in "maintenance" of this image on behalf of other teams, is there any comments DSP @opendatahub-io/data-science-pipelines-maintainers wants to add wrt to this PR?

I can proceed to merge this PR and to my understanding it can be used by the manifests by referring to its main-xxxxxx tag (or sha, if preferred) but @isinyaaa this will still need to be followed-up with DevOps team for the downstream/product, as suggested in #12 (comment)

cc @rareddy
in case I misunderstood anything, kindly let me know!

[diegolovison]

Hi,

I think the commit message should be changed to address the change:
The current commit message: fix missing fields when ordering by
To something like: Adding the columns because of XYZ

This is going to give a notion of the change and the reason behind that.